This conference paper was submitted for presentation at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

Product design in the automotive industry is becoming increasingly demanding as new products should reach high performance standards in very short development cycles. Engineering simulation, using FEA, comes to assist in most product development stages to substitute costly experiments for new designs while speeding up the overall processes. In this direction, optimization procedures are increasingly employed during the design. Apart from FEA, fatigue analysis is a mandatory process which assures product integrity by accurately predicting products'™ life. Using fatigue analysis, the engineer is able to construct stronger yet lighter structures while avoiding overdesign. Special attention should be paid to welded structures since welds often are the weakest part concerning the fatigue life. Fatigue analysis is often incorporated into the product design workflow through an optimization process that fine-tunes structures'™ efficiency. In this case study, a subframe of a car is subjected to a cyclic load. These loads are derived from kinematic model of the car which runs on a Belgian block road for ten seconds. A Multi Body Dynamics solver calculates the dynamics of the entire assembly as well as the stresses and displacements on the subframe which is considered as flexible body. Several design variables are defined directly to the assembly'™s FE model to control its shape and properties. Areas of high stress are selected for shape modification where high damage is likely to occur. Additionally, some non-critical areas are selected which can possibly contribute to mass reduction. Finaly, the seam-welds length is also parametrized, and respective design variables are used. The optimization process is conducted in three steps. Firstly, a DOE (Designs of Experiments) runs to investigate the design variables significance. Then, the created experiments train a Response Surface Model using the Kriging algorithm and finally, optimization runs on the defined Response Surface Model using the Simulated Annealing algorithm. The fatigue analysis is used as a part of the optimization workflow to calculate the damage on critical areas which is the objective to be improved.